Two approaches are presented to improve the capabilities of machine learning models in multi- scale modeling for microstructure homogenization (graphical abstract in Fig. 1). The first approach features a Bayesian data mining scheme with a human in the loop, halving the prediction error com- pared to [1] using four novel efficient to evaluate feature descriptors. The second purely machine learning driven approach utilizes convolutional neural networks, where we introduce a novel mod- ule, designed to capture characteristics of different length scales within the image. The new module features a new normalization block, which aids in calibrating the differently obtained feature char- acteristics. Further improvements, universally applicable to artificial neural networks, are found with a novel hyperparameter insensitive learning rate schedule, which adapts to the training progress of the model. A further improvement is given by a pre-trained feature bypass which utilizes global low level features, to serve as baseline prediction such that the model is able to dedicate its atten- tion to high level features. The proposed schemes have been applied to different literature models, yielding significant improvements in any investigated convolutional neural network. The improve- ments found by the two overarching contributions, i.e., derived through feature development with a human in the loop, and via convolutional neural networks, are critically assessed in a thermal and mechanical setting. It is further expanded to variable material parameters while allowing for variable microstructural elements, yielding drastically reduced prediction errors across-the-board.